Re: Sleep/Lenovo SL410 fails again after csup & clang

  On 10/20/10 11:46, Jung-uk Kim wrote:
> On Wednesday 20 October 2010 01:52 pm, Matt wrote:
>>    On 10/20/10 10:37, Matt wrote:
>>>> On Wed, 20 Oct 2010, Jung-uk Kim wrote:
>>>>> On Tuesday 19 October 2010 09:15 pm, Matt wrote:
>>>>>>    My experience with a sleeping freebsd laptop has been
>>>>>> shortlived!
>>>>>>
>>>>>> Today I rebuilt world using clang&  this morning's csup
>>>>>> current. Clang build went just swimmingly.
>>>>>>
>>>>>> Unthinkingly, I closed my laptop lid and put it in my case.
>>>>>> When I got to my house, it was roasting with fans spinning and
>>>>>> sleep light flashing. No damage, thankfully.
>>>>>>
>>>>>> Low and behold, hw.pci.do_power_resume=0 no longer lets my
>>>>>> laptop sleep!
>>>>>>
>>>>>> I had recently fiddled with powerd, but problem persisted
>>>>>> after reverting to previous configuration of associated
>>>>>> sysctls etc.
>>>>>>
>>>>>> Interestingly, sleep bounce now fails with a hard freeze,
>>>>>> which it never has in the past.
>>>>>>
>>>>>> Verbose output shows the wifi then re0 network interfaces
>>>>>> going from D0->D3 as last living output.
>>>>>>
>>>>>> Please note problem persists regardless of user, X running,
>>>>>> sleep_delay sysctl, do_power_resume, do_power_nodriver, powerd
>>>>>> running/not running.
>>>>>>
>>>>>> Without sleep bounce, problem is characterized by flashing
>>>>>> sleep light and spinning fans (CPU temperature is high).
>>>>>>
>>>>>> No devices added or removed, was sleeping this morning before
>>>>>> buildworld. Is it worth rebuilding with gcc? Or a content
>>>>>> change and not a compiler issue? Any major pci changes lately
>>>>>> maybe?
>>>>> Can you please update source and try again?  If it does not
>>>>> work, please set a new tunable "hw.pci.do_power_suspend=0" and
>>>>> tell me whether it helps or not.  FYI, hw.pci.do_power_resume
>>>>> does not apply to suspend any more.  So, if you want to restore
>>>>> the previous behaviour, you need both
>>>>> "hw.pci.do_power_resume=0" and "hw.pci.do_power_suspend=0".
>>>>> However, my hunch tells me that you only need the latter.
>>>>>
>>>>> Also, if possible, I'd like to see 'devinfo -rv', 'pciconf
>>>>> -clv', 'acpidump -dt', and 'lspci -vvv' output.  Note lspci is
>>>>> available via ports/sysutils/pciutils.
>>>>>
>>>>> Thanks,
>>>>>
>>>>> Jung-uk Kim
>>>>> _______________________________________________
>>>>> freebsd-current_at_freebsd.org mailing list
>>>>> http://lists.freebsd.org/mailman/listinfo/freebsd-current
>>>>> To unsubscribe, send any mail to
>>>>> "freebsd-current-unsubscribe_at_freebsd.org"
>>> A>>  dump file of the above commands in order:
>>> http://pastebin.com/KhFn9xaK
>>>
>>> I'm grabbing a new csup now, glad it wasn't clang at least. When
>>> I did Mac development, LLVM binaries were generally significantly
>>> faster for some things.
>>>
>>> Interesting, I got some phone related topic added into your post,
>>> and no direct mail, so hopefully you can get your hands on the
>>> pastebin link.
>>>
>>> Thank you, I will let you know if the new tunable does the trick.
>>>
>>> Matt
>> Build fails. I will also try gcc, but it seems like it doesn't know
>> about do_power_suspend yet!
>>
>>
>> cd /usr/obj/usr/src/sys/ONOSENDAI; MAKEOBJDIRPREFIX=/usr/obj
>> MACHINE_ARCH=amd64  MACHINE=amd64  CPUTYPE=nocona
>> GROFF_BIN_PATH=/usr/obj/usr/src/tmp/legacy/usr/bin
>> GROFF_FONT_PATH=/usr/obj/usr/src/tmp/legacy/usr/share/groff_font
>> GROFF_TMAC_PATH=/usr/obj/usr/src/tmp/legacy/usr/share/tmac
>> _SHLIBDIRPREFIX=/usr/obj/usr/src/tmp  VERSION="FreeBSD 9.0-CURRENT
>> amd64 900023"  INSTALL="sh /usr/src/tools/install.sh"
>> PATH=/usr/obj/usr/src/tmp/legacy/usr/sbin:/usr/obj/usr/src/tmp/lega
>> cy/usr/bin:/usr/obj/usr/src/tmp/legacy/usr/games:/usr/obj/usr/src/tm
>> p/usr/sbin:/usr/obj/usr/src/tmp/usr/bin:/usr/obj/usr/src/tmp/usr/gam
>> es:/sbin:/bin:/usr/sbin:/usr/bin make KERNEL=kernel all
>> -DNO_MODULES_OBJ
>> clang -c -O2 -fno-strict-aliasing -pipe -march=nocona -std=c99
>> -Wall -Wredundant-decls -Wnested-externs -Wstrict-prototypes
>> -Wmissing-prototypes -Wpointer-arith -Winline -Wcast-qual  -Wundef
>> -Wno-pointer-sign -fformat-extensions -nostdinc  -I. -I/usr/src/sys
>> -I/usr/src/sys/contrib/altq -D_KERNEL -DHAVE_KERNEL_OPTION_HEADERS
>> -include opt_global.h  -fno-omit-frame-pointer -mcmodel=kernel
>> -mno-red-zone  -mfpmath=387 -mno-sse -mno-sse2 -mno-sse3 -mno-mmx
>> -mno-3dnow  -msoft-float -fno-asynchronous-unwind-tables
>> -ffreestanding -fstack-protector   /usr/src/sys/dev/pci/pci_pci.c
>> clang: warning: argument unused during compilation: '-mfpmath=387'
>> /usr/src/sys/dev/pci/pci_pci.c:450:20: error: use of undeclared
>> identifier 'pci_do_power_suspend'; did you mean
>> 'pci_do_power_resume'? if (error == 0&&  pci_do_power_suspend) {
>>                             ^~~~~~~~~~~~~~~~~~~~
>>                             pci_do_power_resume
>> In file included from /usr/src/sys/dev/pci/pci_pci.c:51:
>> /usr/src/sys/dev/pci/pci_private.h:41:13: note:
>> 'pci_do_power_resume' declared
>>         here
>> extern int      pci_do_power_resume;
>>                   ^
>> 1 error generated.
>> *** Error code 1
>>
>> Stop in /usr/obj/usr/src/sys/ONOSENDAI.
>> *** Error code 1
>>
>> Stop in /usr/src.
>> *** Error code 1
>>
>> Stop in /usr/src.
> It seems your sys/dev/pci/pci_private.h is out of sync somehow.  It
> was updated with pci_pci.c in the same commit.  Please try csup
> again.
>
> Jung-uk Kim
>
After current cvsup, SL410 sleeps again (with both 
hw.pci.do_power_suspend and hw.pci.do_power_resume set to 0). Initial 
toggling indicated that at least would work, I will experiment more. 
Thank you all for a quick resolution to the problem!

Matt